One warehouse for ads, checkout and cohorts,
and real unit economics per channel
Ad spend sits in one dashboard. Checkout numbers sit in a second. Completion or retention sits nowhere at all. That makes it impossible to say which channel actually produces students who finish and come back. We connect it all into one warehouse and give you the real unit economics per cohort and channel.
Why the winning channel is a guess
A typical course or coaching business tracks ad spend in the ad platform’s own dashboard. Checkout and revenue live in the payment processor. Enrollment or completion lives in the course platform. Three systems that were never designed to agree with each other. Platform-reported attribution for digital products commonly overstates or understates true conversion by 20 to 40 percent. That gap shows up once cross-device behavior, ad blockers and multi-touch journeys are accounted for. That is an industry pattern, not a flaw unique to any one platform. It means a founder deciding where to spend the next ad dollar is often deciding on numbers that do not reflect reality.
Retention and completion are the deeper gap. They almost never make it into any dashboard at all. A channel that produces cheap enrollments but low completion and high refund rates can look like the best performer in an ads dashboard. It can actually be destroying margin, once support time, refunds and reputation are counted. Without a single warehouse joining spend to checkout to what actually happens inside the course, a business optimizes for the wrong number without knowing it. The decision that matters most, which channel and which offer to scale, gets made on the metric that is easiest to see, not the one that is correct.
Cohort-level blindness compounds the problem. A course business looking only at lifetime totals, total revenue, total refunds, total completions, averages away exactly the signal that matters. That signal is whether the offer, the price point or the audience changed between one launch and the next. A founder comparing this quarter’s launch to last quarter’s on gut feeling alone is really comparing two different audiences and two different ad accounts. Sometimes two different offers too. None of the underlying numbers are broken out cleanly enough to say which change caused which result.
What the warehouse actually tracks
A warehouse joining ad spend, checkout and course data. Meta, Google and TikTok Ads on the spend side. Stripe or your payment processor for checkout and refunds. Kajabi, Teachable, Thinkific or a custom platform for enrollment, progress and completion. All reconciled into one PostgreSQL database with a daily sync and backfill, so history is not lost.
A full funnel from click to completion. Ad click, landing page, checkout, module progress and course completion, mapped as one funnel instead of three disconnected numbers. You see exactly where a cohort’s attention or money is being lost.
Unit economics per cohort, channel and offer. CAC, LTV, refund rate and completion rate, computed per cohort and per channel rather than averaged across your whole history. A launch two months ago and the one running now rarely look the same.
Dashboards built around the decisions you actually make. Which channel to scale. Which offer to retire. Which cohort’s completion rate signals a content problem rather than a marketing one.
An AI analyst for ongoing questions. A read-only role on your warehouse with a guarded SQL layer. Ask it in Telegram which channel brought the highest-completion cohort last quarter, and it answers with the query it ran, instead of you waiting for a report.
Alerts for the numbers that need a fast reaction. Spend spiking against a stale budget. Checkout conversion dropping mid-launch. A refund rate climbing on a specific cohort. Each triggers an alert instead of waiting to be noticed in a weekly review, which matters most during the exact window a launch is running.
How the build runs
- Audit. What exists across your ad accounts, checkout and course platform, what is tracked correctly, and what is wrong or missing. About one week.
- Model. Which questions the business must be able to answer. That drives the design of the data marts, not the other way around.
- Connectors and sync. Raw data pulled from each platform with a daily job and historical backfill, reconciled against what you actually see in your bank account.
- Dashboards and alerts. Built with the people who will actually use them, with alerts to Telegram for anomalies in spend, conversion or refund rate.
- AI analyst. A read-only role, a guarded SQL layer, tested on a set of real recorded questions before being trusted with live data.
What it costs
| Package | Price | Best for |
|---|---|---|
| Tracking audit | from $800 | Finding out what your current tracking actually sees and misses, with a fix list |
| Warehouse and dashboards | from $2,500 | Ad spend, checkout and course platform in one database, with cohort and channel economics |
| AI analyst and monitoring | from $5,000 | Everything in the warehouse package plus an AI analyst in Telegram and anomaly alerts |
Final price depends on how many platforms you connect and how far back the history goes.
What the numbers usually look like
Course businesses that build a proper warehouse typically discover that platform-reported attribution missed 20 to 40 percent of real enrollments, or misattributed them to the wrong channel. That is an industry pattern, not a number specific to any one case, and it routinely changes which channel looks like the winner. Unit economics computed per cohort instead of averaged commonly reveal that one or two cohorts were unprofitable, once refunds and support time are counted. That signal is invisible in a simple revenue-minus-spend view.
Completion and retention data, once connected to acquisition channel, typically shows that the cheapest channel is not always the one producing students who finish and refer others. Refund rate, tracked per channel and per offer rather than as one blended figure, commonly turns out to concentrate heavily in a single segment. That is a far more actionable finding than an average that hides it. Our own numbers are in the case study linked below.
Why Senator Media
We build this warehouse the same way we build one for our own businesses. The questions come first, the data model follows. The AI analyst only gets access to a mart we have already audited for mistakes, never raw platform exports. Pricing is fixed before work starts, not billed by the hour. You see a working connector or dashboard every week during the build, not one reveal at the end.
After launch, the warehouse keeps syncing daily, and we stay available for tuning as your funnel and offers change. We have built unit-economics models for other businesses where the honest answer was that the numbers did not work yet. Reporting that clearly, with the specific lever that would fix it, is the same standard we hold a course business’s numbers to.
If the funnel itself, not just the numbers behind it, needs work, the AI agent for online courses handles sales qualification and checkout recovery on the same data. The full analytics and data service covers everything else we build the same way. For an example of a warehouse joining ad platforms, a store and a CRM into one picture, see the analytics hub built for two brands.
Tell us which platforms you run ads, checkout and your course on. We will send back a fixed price and a three-to-six-week plan: get in touch.
FAQ
How much does this cost and what do we get for it?
From $2,500 for a warehouse connecting your ad platforms, checkout and course platform, with dashboards covering spend, conversion and cohort retention. An AI analyst and ongoing monitoring layer on top is $5,000 and up. Want to know what is broken first? A simpler tracking audit alone starts around $800.
How long does it take?
3 to 6 weeks for the warehouse and dashboards. Adding the AI analyst and monitoring extends that to 6 to 10 weeks. You see working pieces every week rather than waiting for one big reveal: a connector live, then a first dashboard.
Which ad platforms, checkout and course platforms do you connect?
Meta Ads, Google Ads and TikTok Ads on the spend side. Stripe, PayPal or your payment processor for checkout. Kajabi, Teachable, Thinkific or a custom platform for enrollment and completion. Plus GA4 or PostHog for site behavior.
Can an AI analyst really be trusted with real business numbers?
It runs read-only SQL on a prepared data mart, one query at a time, with forced limits. It shows the query it ran rather than giving a black-box answer. We audit our own SQL guard before trusting it with live data. The answer is only as good as the mart, which is why the mart comes first.
We are a solo coach, not a big course business. Is this overkill?
Start with a tracking audit rather than the full warehouse. For a business running one or two cohorts a year, the audit alone usually surfaces where attribution is lying, before you spend on the bigger build.
Can it show economics per cohort, not just overall?
Yes, that is the main point. A launch two cohorts ago and the one running now can have very different CAC, completion and refund rates. Averaging them together hides exactly the signal you need to decide what to change next.